Terrific, thank you!
Jon_Behar
Karma: 1,349
Thanks David, that all makes sense. For future iterations of this analysis, I’d be strongly in favor of adding a sentence about Peter like you had in last year’s summary.
Personal Contacts (14%), LessWrong (9.6%) and 80,000 Hours (9.6%) are still the main ways most people have heard of EA over time.
Shouldn’t Peter Singer be on this list? He showed up in 203 of the open ended responses (9.5% of 2137 total responses), and that doesn’t count any open ended comments that didn’t mention him by name or non-open ended responses that he’s associated with (e.g. the people who answered they heard about EA through TLYCS the organization, which I work for in the interest of disclosure).
Thank you for conducting and sharing this analysis!
I like the book distribution idea, but would suggest using the updated 10th Anniversary edition of The Life You Can Save. While its focus on global poverty makes it narrower than Doing Good Better, it has several advantages:
· The audiobook and ebook are available for free
· The celebrity narrators (e.g. Kristen Bell and Paul Simon) add credibility
· It’s more up to date
That said, I think it’d be really interesting to do an experiment comparing the efficacy of distributing each book.
Note: I work for TLYCS the organization
The “Worm Wars” could arguably be an example (though the contentious research was not just from the EA community)
Most of the people on The Life You Can Save’s team have significant experience in the for-profit sector, which I think is relatively rare in the EA community. Charlie Bresler, our Executive Director, used to be the President of the Men’s Wearhouse. And before I served as COO for an extended period, I spent ~10 years in the finance sector at Bridgewater Associates. So I think those experiences helped shape The Life You Can Save’s culture. For instance, I think due to the diversity of the backgrounds of our team members, we may engage with a significantly more diverse range of stakeholders on a day-to-day basis than many “typical” organizations in this field. The advantage of this is that it provides us with a variation and depth of expertise to draw upon when we are making strategic decisions about our organization’s mission and approach.
Additionally, our team is also, generally speaking, older than most of the EA community. I’m not sure I can point to specific things that causes us to know about, unless I go with a tongue-in-cheek answer like “what the 1980s were like”.
Here’s how our Oxfam information page describes why we recommend them. (FYI, on each charity’s information page we have an FAQ explaining our recommendation).
We recommend Oxfam for donors who want to support a large, multinational organization working to fight global poverty in a wide variety of ways and in a wide variety of places. Because Oxfam is so large, including them as a recommended charity also significantly expands the tax-deductible giving options and program locations we can offer our global audience.
Due to the breadth and scope of its work, Oxfam’s impact is inherently difficult to measure and attribute, but we believe that some hard-to-measure work, like advocacy, can be a powerful way to help people in extreme poverty. Full disclosure: Peter Singer is a member of Oxfam America’s Leadership Council, an unpaid advisory body of significant Oxfam donors. Peter’s wife, Renata Singer, was employed by Oxfam Australia in the 1990s as their publications officer.
The fact that we don’t recommend The End Fund definitely shouldn’t be interpreted as a negative assessment of their work. Rather, it relates to your other question about the “paradox of choice.” We recommend SCI and Evidence Action (which runs Deworm the World), and generally don’t want to recommend many charities performing similar interventions without a compelling reason, as we think this will be confusing to donors.
In some cases, we do think there’s a good reason to have multiple charities performing similar interventions. For instance, we added Malaria Consortium to our list (which already included AMF) when GiveWell rated the former’s marginal cost-effectiveness as higher than the latter’s. We also have multiple food fortification recommendations which were added at the same time, and which we didn’t feel like we had good reason to distinguish between, but once those were already on our list we declined to add the Food Fortification Initiative when GiveWell later added it as a standout charity.
As the previous examples show, there’s some path dependency to our list (i.e. the order in which we add charities matters). This reflects our belief that 1) all else equal, we want our list to be simple for donors with minimal overlap and 2) we think removing a charity from our list because we added a similar one that might be slightly better would send an inappropriately negative signal about the charity we removed.
We’ve worked with HNW donors to determine which of our recommended organizations are the best match for the donors’ specific values and causes of interest. So far we haven’t done any bespoke research for donors, though this is definitely an area we expect to expand into in the future.
We’ll sometimes get inquiries about causes outside our scope. Where possible, we refer them to EA resources, such as ACE for animal welfare and Founders Pledge’s research on climate change. (We also have links to those two organizations at the bottom of our charity selection methodology writeup in a “beyond global poverty” section).
We recommend a significantly broader set of charities than GiveWell, which is an intentional strategy to offer donors a wider range of options. That said, in the near-term any additions to our list are likely to come from GiveWell. We had been sourcing new recommendations from Impact Matters as well, but they’ve recently pivoted away from the in-depth “impact audits” we’d been relying on and toward much shallower reviews of many more charities. We’ve written up our selection process in more detail here.
Down the road, we’d like to add dedicated staff to work on charity assessment. The primary obstacle to this is lack of funding. We expect this staff would curate research from GiveWell and other sources more than doing primary research. We think dedicated staff would be helpful in expanding our list into cause areas where there’s donor demand (e.g. education and climate change), developing an overall fund and funds for specific cause areas, and offering more concierge services to high net worth donors.
This is something that’s definitely on our radar screen due to climate change’s outsized impact on people living in extreme poverty. We also think it’s a cause area that’s of interest to donors (both our existing donor base and others). However, it’s unlikely that we’ll move forward on this until we have the capacity to add dedicated staff for charity assessment.
These aren’t really surprises, but the experience reinforced a couple of things: celebrities are really busy, and they have a huge reach (e.g. 1 instagram post from Kristen Bell led to over 1000 people downloading the book and subscribing to our newsletter.)
At least a few of the celebrities seem interested beyond seeing us as a random good cause. As an example, Michael Schur really engaged with the intellectual substance of the book in the foreword he wrote for the new edition. And that probably shouldn’t be surprising, as his show The Good Place is essentially oriented around some similar themes.
1. Agree that equity valuations outside the US are much less extreme. But if you’re building a diversified portfolio, global fixed income and US equities are probably going to play a large part. So avoiding lower expected returns in those asset classes would require an element of active management, which I think raises the hurdle for this project significantly since active management is both expensive and hard to do well. Given the goals of this fund, I would think a passive Risk Parity strategy (which includes a lot of fixed income) would make a lot of sense.
2. Good point. I would still argue that if there’s an intention to deploy money when “market changes make investing much less attractive”, it makes sense to try and define those types of conditions ahead of time. And if you were going through that exercise a couple of hundred years ago, I’m pretty sure “widespread negative real yields” would have made the list.
Money would only be deployed when there is a strong case that allocating to a funding opportunity is higher-impact from a longtermist perspective than keeping the money invested. This could happen, for instance, if our estimate of the expropriation rate rises greatly, legal and/or market changes make investing much less attractive, or we identify a truly extraordinary funding opportunity that we don’t expect to be filled by others.
I think this fund is an intriguing idea. But I think there’s an argument that current market conditions would suggest deploying funds now rather than investing them. Interest rates are at extraordinarily low levels, which suggest lower than normal expected future returns not only for fixed income instruments but also other asset classes that compete for capital with fixed income. To put into context how low current rates are, a 2015 analysis found that “rates remain at the lowest levels in the last 5,000 years of civilization.” Since then, rates have gone even lower. As of August 2019, “About $15 trillion of government bonds worldwide, or 25% of the market, now trade at negative yields, according to Deutsche Bank.”
To be clear, I think this argument applies to the general class of “donate now vs. invest and donate later decisions” EAs make, not just the proposed longtermist fund (where it might apply less due to the especially long time horizons). But my impression is that EAs are often too quick to assume they can always achieve investment returns in line with historical long-term averages, when they should only expect to do so over very long time horizons or when starting valuations are also in line with historical long-term averages.
Of course, it’s also worth noting that the “valuations are currently high so on the margin EAs should give more now” argument could have been made several years ago, and those years have generally been good ones for asset prices…
Peter Singer on the 80K podcast
Thanks to everyone who’s helped share and promote the book launch! It’s very early, but so far the numbers are looking good. Here’s an excerpt of an email I sent to my colleagues at TLYCS looking at metrics for the book launch on Giving Tuesday:
Here are the highlights:
Money Moved up 100%: We moved $16,477 through Network for Good (online donations through our site, excluding Australia) on Giving Tuesday (GT), up almost exactly 2x vs. $8244 from GT last year. The real picture is even better than the headline numbers, as we’ve seen a substantial improvement in recurring giving, which the new site really emphasizes. On GT18, we had only $340 in recurring gifts. This year, we had $1332 up almost 4x. (In both years, recurring gifts were almost entirely monthly).
Very strong subscription growth: We’ve gotten ~1500 subscriptions through the new site, almost all of which have come via the book download. We now have ~16,500 total subscribers, so this is a big bump for us: so far this week, we’ve increased our newsletter audience by 10%!
~2000 Book Downloads: We’re still nailing down some aspects of our download tracking, so please take these numbers with a grain of salt. Downloads are also our hardest metric to contextualize; we don’t have a great benchmark for our performance. That said, we’ve seen about ~2000 book downloads, split roughly evenly between the e-book and audio-book. Those downloads come from roughly ~2600 people who submit the download form (people can complete the form multiple times), so we may see more downloads as more people follow the instructions in the emails they’ve received (though we should also expect some attrition). Audiobook downloads are split roughly in half between people who download the full book or just a chapter, meaning roughly ¼ of downloads are partial downloads of one or more chapters. (Important but not urgent: we should consider ways to nudge people more toward the full book, and ways to get people who download 1 chapter to download more in the future.)
Web Traffic: We had over 6000 visitors on GT, up ~2x vs. 2600 on GT 2018, 3600 on 12/3/18, and an average of ~2800/day throughout November 2019. Traffic looks good across all our channels. The one exception is Referral Traffic which looks quite low, meaning we haven’t been getting a lot of traffic from links on other peoples sites so far. But we’ve had large increases in visitors from social media, our email campaigns, and direct traffic (people typing the URL directly) the latter of which could be related to a lot of things including some podcast ads that started yesterday. We should expect to see more referral traffic as we get deeper into giving season.
I like this additional categorization, the “community infrastructure” distinction seems valuable. Thanks for sharing!
Balanced budget amendments and/or debt limits
The Life You Can Save is working with an app-development company called Meepo (which is doing pro bono work) to build a non-profit donation app, which is currently in beta. You can learn more about this project, and how to download the beta version, here.
The Life You Can Save is working with an app-development company called Meepo (which is doing pro bono work) to build a non-profit donation app, which is currently in beta. You can learn more about this project, and how to download the beta version, here.
Free distribution seems to have helped a lot. The original version sold ~45,000 copies in its first 10 years; in its first 6 months, we’ve distributed roughly the same number of copies of the 10th anniversary edition.
The original edition has presumably had more readers than the updated version so far: over 10 years you can rack up a lot of library checkouts and used book readings that aren’t captured in the sales numbers, and people are more likely to consume a book if they’ve paid for it than if they got it for free. But based on the first 6 months, I’d expect the new edition to be read many more times over the long term, and for that to be driven by it being freely available. (I’d also expect factors like promotion by the celebrity narrators of the audiobook to increase distribution.)